arXiv:math/0602560v1 [math.AP] 24 Feb 2006
GLOBAL WELL-POSEDNESS FOR A PERIODIC NONLINEAR
SCHRÖDINGER EQUATION IN 1D AND 2D
DANIELA DE SILVA, NATAŠA PAVLOVIĆ, GIGLIOLA STAFFILANI, AND NIKOLAOS TZIRAKIS
Abstract. The initial value problem for the L2 critical semilinear Schrödinger equation
with periodic boundary data is considered. We show that the problem is globally well
posed in H s (Td ), for s > 4/9 and s > 2/3 in 1D and 2D respectively, confirming in 2D a
statement of Bourgain in [3]. We use the “I-method”. This method allows one to introduce
a modification of the energy functional that is well defined for initial data below the H 1 (Td )
threshold. The main ingredient in the proof is a “refinement” of the Strichartz’s estimates
that hold true for solutions defined on the rescaled space, Tdλ = Rd /λZd , d = 1, 2.
1. Introduction
In this paper we study the L2 critical Cauchy problem
4
(1.1)
iut + ∆u − |u| d u = 0, x ∈ Td , t ≥ 0
(1.2)
u(x, 0) = u0 (x) ∈ H s (Td ),
where d = 1, 2 and Td = Rd /Zd is the d-dimensional torus.
We say that a Cauchy problem is locally well-posed in H s if for any choice of initial
data u0 ∈ H s , there exists a positive time T = T (ku0 kH s ) depending only on the norm
of the initial data, such that a solution to the initial value problem exists on the time
interval [0, T ], is unique and the solution map from Hxs to Ct0 Hxs depends continuously on
the initial data on the time interval [0, T ]. If T = ∞ we say that a Cauchy problem is
globally well-posed.
In the case when x ∈ Rd local well-posedness for (1.1)-(1.2) in H s (Rd ) has been studied
extensively (see, for example, [8, 16, 20]). In particular if one solves the equivalent integral
equation by Picard’s fixed point method and controls the nonlinearity in the iteration
process by using Strichartz’s type inequalities, then the problem can be shown to be locally
well-posed for all s > 0.
Bourgain [1] adjusted this approach to the periodic case, where there are certain difficulties due mainly to a “lack of dispersion”. In [1] number theoretic methods were used to
show that (1.1)-(1.2) is locally well-posed in H s (Td ), d = 1, 2 for any s > 0.
Assuming local existence there are many issues to be addressed about the behavior of
the solution as t → ∞,
• global well-posedness/blow-up behavior
• asymptotic stability
• behavior of higher order Sobolev norms of smooth solutions.
Date: February 16, 2006.
1
2
DANIELA DE SILVA, NATAŠA PAVLOVIĆ, GIGLIOLA STAFFILANI, AND NIKOLAOS TZIRAKIS
In this note we address the question of global well-posedness for (1.1)-(1.2). We recall
that solutions of (1.1) satisfy mass conservation
ku(t)kL2 = ku0 kL2
and smooth solutions also satisfy energy conservation
Z
Z
4
d
1
2
|∇u(t)| dx +
|u(t)|2+ d dx = E(u0 ).
E(u)(t) =
2
2(d + 2)
For initial data in H s (Td ) , s ≥ 1, the energy conservation together with local well-posedness
imply global well-posedness for s ≥ 1. However it is a more subtle problem to extend the
global theory to infinite energy initial data. Bourgain [3] established global well posedness
for (1.1) in H s (T) for any s > 1/2−, by combining a “normal form” reduction method (see
also [5]), the “I-method”, and a refined trilinear Strichartz type inequality. The normal
form reduction is achieved by symplectic transformations that, in some sense, reduce the
nonlinear part of the equation to its “essential part”. The I-method, introduced by J.
Colliander, M. Keel, G. Staffilani, H. Takaoka and T. Tao in [10, 13, 14], is based on the
almost conservation of certain modified Hamiltonians. These two methods together yield
global well-posedness in H s (T) with s ≥ 1/2. The refined trilinear Strichartz inequality
established in [3] is a qualitative estimate needed to achieve global well-posedness in H s (T),
s∗ < s < 1/2 for some s∗ .
In this paper we continue to fill in the gap between what is known locally and what is
known globally in H s (Td ) for s > 0, when d = 1, 2. Our approach is based on an implementation of the I-method itself adjusted to the periodic setting via elementary number
theoretic techniques.
In order to present our method, we briefly review global well-posedness results on Rd .
Using an approximation of the modified energy in the I-method, Tzirakis [25] showed that
the Cauchy problem (1.1)-(1.2) is globally well posed in H s (R) for any s > 4/9. For
u0 ∈ H s (R2 ) the best known global well posedness result for (1.1)-(1.2) is in [14] where
the authors proved global well posedness for any s > 4/7. In both results, apart from the
application of the I-method, a key element in the proof is the existence of a bilinear refined
Strichartz’s estimate due to Bourgain, [4] (see also [9] [10]). In general dimensions d ≥ 2
this estimate reads as follows. Let f and g be any two Schwartz functions whose Fourier
transforms are supported in |k| ∼ N1 and |k| ∼ N2 respectively. Then we have
d−1
kUt f Ut gkL2t L2x (R×Rd ) ≤ C
N2 2
1
2
kf kL2 (Rd ) kgkL2 (Rd ) ,
N1
where Ut denotes the solution operator associated to the linear Schrödinger equation. In
one dimension the estimate fails if the two frequencies are comparable but continues to hold
if the frequencies are separated (N1 >> N2 ). Such an estimate is very useful when f is in
high frequency and g is in low frequency since we can move derivatives freely to the low
frequency factor.
The main difficulty in obtaining global well-posedness results in the periodic context is
exactly the absence of a quantitative refinement of the bilinear Strichartz’s estimate. The
reader can consult the paper of Kenig, Ponce and Vega, [21], where the difference between
the real and the periodic case is clearly exposed when one tries to prove bilinear estimates
in different functional spaces.
In order to overcome the non-availability of a quantitative refined Strichartz’s estimate,
we develop a different strategy, closer in spirit to the approach in [13]. By exploiting the
GLOBAL WELL-POSEDNESS FOR A PERIODIC NONLINEAR SCHRÖDINGER EQUATION
3
scaling symmetry of the equation, we analyze (1.1) on Tdλ ×R, d = 1, 2, where Tdλ = Rd /λZd .
The main novelty in this analysis consists in establishing a bilinear Strichartz’s inequality
for λ-periodic functions which are well separated in frequency space. The constant in the
inequality is quantified in terms of λ. As λ → ∞, our estimate reduces to the refined
bilinear Strichartz’s inequality1 on Rd , d = 1, 2, see, for example, [4, 10, 23]. Such an
estimate allows us to use the I-method machinery in an efficient way.
More precisely, when d = 1, beside rescaling, we follow the argument in [25] (see also
[12]), where one applies the I-operator to the equation on Tλ , and defines a modified second
energy functional as the energy corresponding to the new “I-system”. We prove that such
modified second energy is “almost conserved”, that is the time derivative of this new energy
decays with respect to a very large parameter N . Roughly speaking, N denotes the stage at
which the I-operator stops behaving like the identity and starts smoothing out the solution.
We prove a local existence result for the “I-system”, under a smallness assumption for the
initial data, which is guaranteed by choosing λ in terms of N . The decay of the modified
energy enables us to iterate the local existence preserving the same bound for kuλ kH s during
the iteration process. By undoing the scaling we obtain polynomial in time bounds for u
in T × R+ , and this immediately implies global well posedness.
The precise statement of our 1D result reads as follows.
Theorem 1.1. The initial value problem (1.1)-(1.2) is globally well-posed in H s (T) for
s > 49 .
For the two dimensional case Bourgain already announced in [3] that the I-method
based only on the first energy would give global well-posedness for s > 2/3. While we
were explicitly writing up the calculations to recover this claim, we noticed that in one
particular case2 of the estimate of the first energy, a better Strichartz inequality was needed
to successfully conclude the argument. We then proceeded by determining a qualitative
“ǫ-refined” Strichartz type estimate, see Proposition 4.6, which allowed us to implement
the I-method described above to obtain indeed the following result:
Theorem 1.2. The initial value problem (1.1)-(1.2) is globally well-posed in H s (T2 ) for
s > 32 .
We remark that, as we mentioned above, in 1D we introduce an approximation to the
modified energy, by adding correction terms, in the spirit of [25] and [12]. In the 2D
problem we did not use correction terms, since the Fourier multipliers corresponding to
the approximated modified energy would be singular. This singularity is not only caused
by the presence of zero frequencies, but also by orthogonality issues. This is a whole new
ground that we are exploring in a different paper.
Organization of the paper. In section 2 we review the notation and some known
Strichartz estimates for the periodic Schrodinger equations in 1D and 2D. In section 3
we present the proof of Theorem 1.1, while in section 4 we give a proof of Theorem 1.2.
Acknowledgments. We would like to thank Akshay Venkatesh and Terence Tao for valuable discussions. N.P. received partial support from NSF grant DMS-0304594. N.T. received partial support from NSF grant DMS-0111298. Portions of this work were carried
out at the Institute for Advanced Study (IAS) in Princeton and at the Mathematical
1In a way we are introducing “more dispersion” by rescaling the problem.
2See Case IIIb) in the proof of Proposition 4.7.
4
DANIELA DE SILVA, NATAŠA PAVLOVIĆ, GIGLIOLA STAFFILANI, AND NIKOLAOS TZIRAKIS
Sciences Research Institute (MSRI) at Berkeley. The authors would like to thank these
institutions for their hospitality.
2. Notation
In what follows we use A . B to denote an estimate of the form A ≤ CB for some
constant C. If A . B and B . A we say that A ∼ B. We write A ≪ B to denote an
estimate of the form A ≤ cB for some small constant c > 0. In addition hai := 1 + |a| and
a± := a ± ǫ with 0 < ǫ << 1.
We recall that the equation (1.1) is L2 invariant under the following scaling (x, t) →
1 x t
d
d ( λ , λ2 ). Thus if u(x, t) solves (1.1) on T × R then
λ2
uλ (x, t) =
1
x t
u( , 2 )
λ λ
λ
d
2
is a solution of (1.1) in Td λ × R where Tdλ = Rd /λZd .
Since in our argument we exploit the scaling symmetry let us recall some properties of
λ−periodic functions. Define (dk)λ to be the normalized counting measure on ( λ1 Z)d :
Z
1 X
a(k).
a(k)(dk)λ = d
λ
1
d
k∈( λ Z)
We define the Fourier transform of f (x) ∈ L1x∈[0,λ]d by
Z
e−2πikx f (x)dx.
fˆ(k) =
[0,λ]d
For an appropriate class of functions the following Fourier inversion formula holds:
Z
f (x) = e2πikx fˆ(k)(dk)λ .
Moreover we know that the following identities are true:
(1) kf kL2 ([0,λ]d ) = kfˆkL2 ((dk)λ ) , (Plancherel)
(2)
R
[0,λ]d
f (x)ḡ(x)dx =
(3) fcg(k) = fˆ ⋆λ ĝ(k) =
R
R
¯
fˆ(k)ĝ(k)(dk)
λ , (Parseval)
fˆ(k − k1 )ĝ(k1 )(dk1 )λ ,
We define the Sobolev space H s = H s ([0, λ]d ) as the space equipped with the norm
kf kH s = khkis fˆ(k)kL2 ((dk)λ ) .
We write Uλ (t) for the solution operator to the linear Schrödinger equation
iut − ∆u = 0, x ∈ [0, λ]d ,
that is
Uλ (t)u0 (x) =
Z
2
e2πikx−(2πk) it u
c0 (k)(dk)λ .
We denote by X s,b = X s,b (Tdλ ×R) the completion of S(Tdλ ×R) with respect to the following
norm, see, for example, [15]
GLOBAL WELL-POSEDNESS FOR A PERIODIC NONLINEAR SCHRÖDINGER EQUATION
kukX s,b = kUλ (−t)ukHxs H b = khkis hτ − 4π 2 k2 ib ũ(k, τ )kL2τ L2
t
(dk)λ
5
,
where ũ(k, τ ) is the space-time Fourier Transform
Z Z
e−2πi(k·x+τ t) u(x, t)dxdt.
ũ(k, τ ) =
[0,λ]d
Furthermore for a given time interval J, we define
kf kX s,b =
J
inf kgkX s,b .
on J
g=f
Often we will drop the subscript J.
In what follows we will need the following known estimates. Since
(2πk)2 it
U\
u0 (k)
c
λ (t)u0 (k) = e
we have that
kUλ (t)u0 kL∞
2 . ku0 kL2 .
t Lx
Hence
(2.1)
kukL∞
2 = kUλ (t)Uλ (−t)ukL∞ L2 . kUλ (−t)ukL∞ L2 . kuk 0,1/2+ ,
X
x
x
t Lx
t
t
where in the last inequality we applied the definition of the X s,b spaces, the basic estimate
kukL∞ ≤ kûkL1 , and the Cauchy-Schwartz inequality.
1D estimates. If u is on T1 then the estimate (2.1) combined with the Sobolev embedding
theorem implies that
∞ . kuk 1/2+,1/2+ .
kukL∞
X
t Lx
(2.2)
Also the following linear Strichartz’s estimates was obtained in [1] in the case of the torus:
kukL4t L4x . kukX 0,b ,
(2.3)
for any b > 83 , and
kukL6t L6x . kukX 0+,1/2+ .
(2.4)
We note that (2.2) remains true for the λ-periodic problem. This is the case also for
(2.3). The proof is essentially in [18]. In fact, it is enough to show, by the standard X s,b
method, that
X
1
hτ + k12 + (k − k1 )2 i1−4b . C.
sup
λ (ξ,τ )∈ 1 Z×R
1
λ
k1 ∈ λ Z
This is done in [18], the only difference being that there are O(λ) numbers k ∈ λ1 Z, such
that |k + x0 | < 1 and |k − x0 | < 1, where x20 = |2τ + ξ 2 |. But then summing the above
series, using Cauchy-Schwartz inequality and the fact that b > 3/8, one gets that the left
hand side of the inequality is
1
. (c1 λ + c2 ) . C
λ
for any λ > 1. So from now on we will use (2.3) for the λ-periodic solutions without any
further comment. On the other hand, using scaling we can see that for the λ-periodic
solutions, (2.4) takes the form
(2.5)
kukL6t L6x . λ0+ kukX 0+,1/2+ .
6
DANIELA DE SILVA, NATAŠA PAVLOVIĆ, GIGLIOLA STAFFILANI, AND NIKOLAOS TZIRAKIS
If we interpolate equations (2.2) and (2.5) we get
kukLpt Lpx . λ0+ kukX α1 (p),1/2+ ,
(2.6)
with α1 (p) = ( 12 − p3 )+ and 6 ≤ p ≤ ∞.
2D estimates. On T2 Bourgain proved, [1],
kukL4t L4x . kukX 0+,1/2+ .
(2.7)
Again using scaling we can prove that for the λ-periodic solutions, (2.7) takes the form
kukL4t L4x . λ0+ kukX 0+,1/2+ .
(2.8)
The estimate (2.1) together with Sobolev embedding gives:
∞ . kuk 1+,1/2+ .
kukL∞
X
t Lx
(2.9)
Hence, by interpolation, we get
kukLpt Lpx . λ0+ kukX α2 (p),1/2+ ,
(2.10)
with α2 (p) = (1 − 4p )+ and 4 ≤ p ≤ ∞.
Remark 2.1. (Decomposition remark) Our approach to prove Theorem 1.1 and Theorem
1.2 is based on obtaining certain multilinear estimates in appropriate functional spaces
which are L2 -based. Hence, whenever we perform a Littlewood-Paley decomposition of a
function we shall assume that the Fourier transforms of the Littlewood-Paley pieces are
positive. Moreover, we will ignore the presence of conjugates.
3. The I-method and the proof of Theorem 1.1
In this section we present the proof of Theorem 1.1. We start by recalling the definition
of the operator I introduced by Colliander et al, see, for example, [10, 14]. For s < 1 and
a parameter N >> 1 let m(k) be the restriction to Z of the following smooth monotone
multiplier:
1
if |k| < N
m(k) :=
|k| s−1
if |k| > 2N
(N )
We define the multiplier operator I : H s → H 1 by
c
Iu(k)
= m(k)û(k).
The operator I is smoothing of order 1 − s and we have that:
(3.1)
kukX s0 ,b0 . kIukX s0 +1−s,b0 . N 1−s kukX s0 ,b0
for any s0 , b0 ∈ R.
For u ∈ H s we set
(3.2)
where
E 1 (u) = E(Iu),
Z
Z
1
1
2
E(u)(t) =
|∇u(t)| dx +
|u(t)|6 dx = E(u0 ).
2
6
We refer to E 1 (u) as the first modified energy. As observed by Colliander et al a hierarchy
of modified energies can be formally considered for different nonlinear dispersive equations.
The goal of the I-method is to prove that the modified energies are “almost conserved” i.e.
they decay in time with respect to N . Since in 1D we base our approach on the analysis of
GLOBAL WELL-POSEDNESS FOR A PERIODIC NONLINEAR SCHRÖDINGER EQUATION
7
a second modified energy, it is appropriate to collect some facts concerning the calculus of
multilinear forms used to define the hierarchy, see, for example [25].
If n ≥ 2 is an even integer we define a spatial multiplier of order n to be the function
Mn (k1 , k2 , . . . , kn ) on Γn = {(k1 , k2 , . . . , kn ) ∈ λ1 Zn : k1 + k2 + . . . + kn = 0} which we endow
with the standard measure δ(k1 + k2 + . . . + kn ). If Mn is a multiplier of order n, 1 ≤ j ≤ n
is an index and l ≥ 1 is an even integer we define the elongation Xjl (Mn ) of Mn to be the
multiplier of order n + l given by
Xjl (Mn )(k1 , k2 , . . . , kn+l ) = Mn (k1 , . . . , kj−1 , kj + . . . + kj+l , kj+l+1 , . . . , kn+l ).
In addition if Mn is a multiplier of order n and f1 , f2 , ..., fn are functions on Tλ we define
Z
n
Y
fˆj (kj ),
Mn (k1 , k2 , . . . , kn )
Λn (Mn ; f1 , f2 , ..., fn ) =
Γn
i=1
where we adopt the notation Λn (Mn ; f ) = Λn (Mn ; f, f¯, ..., f, f¯). Observe that Λn (Mn ; f )
is invariant under permutations of the even kj indices, or of the odd kj indices.
If f is a solution of (1.1) the following differentiation law holds for the multilinear forms
Λn (Mn ; f ):
(3.3)
∂t Λn (Mn ) = iΛn (Mn
n
X
j=1
(−1)j kj2 ) +
n
X
iΛn+4 ( (−1)j Xj4 (Mn )).
j=1
Note that in this notation the first modified energy (3.2) reads as:
Z
Z
1
1
1
1
E 1 (u) =
|∂x Iu|2 dx +
|Iu|6 dx = − Λ2 (m1 k1 m2 k2 ) + Λ6 (m1 ...m6 )
2
6
2
6
where mj = m(kj ).
We define the second modified energy
1
1
E 2 (u) = − Λ2 (m1 k1 m2 k2 ) + Λ6 (M6 (k1 , k2 , ..., k6 )),
2
6
where M6 (k1 , k2 , ..., k6 ) is the following multiplier:
(3.4)
M6 (k1 , k2 , ..., k6 ) =
m21 k12 − m22 k22 + m23 k32 − m24 k42 + m25 k52 − m26 k62
.
k12 − k22 + k32 − k42 + k52 − k62
Notice that the zero set of the denominator corresponds to the resonant set for a six-waves
interaction. We remark that M6 contains more “cancellations” than the multiplier m1 ...m6
that appears in E 1 .
The differentiation rule (3.3) together with the fundamental theorem of calculus implies
the following Lemma, which will be used to prove that E 2 is almost conserved.
Lemma 3.1. Let u be an H 1 solution to (1.1). Then for any T ∈ R and δ > 0 we have
Z T +δ
2
2
Λ10 (M10 ; u(t))dt,
(3.5)
E (u(T + δ)) − E (u(T )) =
T
P
with M10 = c {M6 (kabcde , kf , kg , kh , ki , kj ) − M6 (ka , kbcdef , kg , kh , ki , kj ) +
M6 (ka , kb , kcdef g , kh , ki , kj ) − M6 (ka , kb , kc , kdef gh , ki , kj ) +
M6 (ka , kb , kc , kd , kef ghj , kj ) − M6 (ka , kb , kc , kd , ke , kf ghij )},
8
DANIELA DE SILVA, NATAŠA PAVLOVIĆ, GIGLIOLA STAFFILANI, AND NIKOLAOS TZIRAKIS
where the summation runs over all permutations {a, c, e, g, i} = {1, 3, 5, 7, 9} and {b, d, f, h, j}
= {2, 4, 6, 8, 10}. Furthermore if |kj | ≪ N for all j then the multiplier M10 vanishes.
As it was observed in [25] where the equation (1.1) was considered on R one has
Proposition 3.2. The multiplier M6 defined in (3.4) is bounded on its domain of definition.
For the proof see [25].
Now we proceed to present the steps leading to the proof of Theorem 1.1.
Step 1: Local well-posedness for the I-system. The first step towards the proof of
Theorem 1.1 is to apply the I-operator to (1.1) and prove a local well-posedness result for
the I-initial value problem
(3.6)
iIut + Iuxx − I(|u|4 u) = 0
Iu(x, 0) = Iu0 (x) ∈ H 1 (Tλ ), t ∈ R.
In order to obtain the local well-posedness we need the following technical lemma:
Lemma 3.3. Let η ∈ C0∞ be a bump function that is supported on [−2, 2] and equals 1 on
′
′
′
[−1, 1] and set ηδ (t) = η( δt ). For b, b ∈ R with −1/2 < b ≤ 0 ≤ b ≤ b + 1 and δ ≤ 1 we
have:
(1) kη1 (t)Uλ (t)u0 kX s,b . ku0 kH s ,
′
Rδ
(2) kηδ (t) 0 Uλ (t − τ )F (τ )dτ kX s,b . δ1+b −b kF kX s,b′ .
Proof. (1) We compute (see, for example, [15])
kη1 (t)Uλ (t)u0 kX s,b = kUλ (−t)η1 (t)Uλ (t)u0 kH b Hxs
t
= kη1 (t)u0 kH b Hxs
t
= ku0 kH s kη1 (t)kH b
t
∼ ku0 kH s .
(2) The following inequality is true and its proof can be found in [15], [21],
kηδ (t)
Z
δ
0
′
G(τ )dτ kH b . δ1+b
t
−b
kGk
′
Htb
.
But then (2) follows if we apply the above inequality for fixed ξ, multiplying by < ξ >2s
and taking the L2 norm in ξ.
Remark. It is shown in [11] that if
kuvkX s,b−1 . kukX s,b kvkX s,b ,
then
kI(uv)kX 1,b−1 . kIukX 1,b kIvkX 1,b ,
where the constants in the inequality above are independent of N . From now on we use
this fact and refer to it as the “invariant lemma”. For details see Lemma 12.1 in [11].
Now we present the local well-posedness result for (3.6). Because of the loss in the periodic Strichartz estimates (2.5) represented by the λ0+ factor, the local well-posedness is
not as straightforward as in the real case. (There one can essentially use the Leibnitz rule
GLOBAL WELL-POSEDNESS FOR A PERIODIC NONLINEAR SCHRÖDINGER EQUATION
9
for the fractional derivatives and the Cauchy-Schwartz inequality along with interpolated
Strichartz’s estimates to bound the nonlinearity). In 1D we can still follow the same approach thanks to (2.3), which is valid for the λ-periodic problem. In 2D for the λ-periodic
problem we need a more complicated argument that is due to Bourgain, [2], and uses a local
variant of the periodic Strichartz inequality that is proved there. For details see [2]. As a
final comment, note that our local well-posedness results are not optimal but they suffice
for the purposes of the global theory that we establish. Thus without aiming at sharpness,
we prove local well-posedness for s > 5/16 in 1D (see Proposition 3.4 and Corollary 3.5)
and s > 7/20 in 2D (see Proposition 4.1 and Corollary 4.2).
Proposition 3.4. Consider the I-initial value problem (3.6), λ = 1. If kIu0 kH 1 . 1, then
5
in [0, δ] ∼ [0, 1].
(3.6) is locally well-posed for any s > 16
Proof. By Duhamel’s formula and Lemma 3.3 we have that
1
kIukX 1,1/2+ . kIu0 kH 1 + δ 2 −ǫ kI(|u|4 u)kX 1,−1/2+2ǫ .
(3.7)
We shall prove that
kI(|u|4 u)kX 1,−1/2+2ǫ . kIuk5X 1,1/2+ .
By the invariant lemma we know that to prove such an estimate it suffices to prove
k|u|4 ukX s,−1/2+2ǫ . kuk5X s,1/2+ .
By Hölder’s inequality combined with the Leibnitz rule ([22]) we have
k|u|4 ukX s,−1/2+2ǫ ≤ k|u|4 ukX s,0 . kJ s ukL4t L4x kuk4L16 L16 ,
t
x
where J s is the Bessel potential of order s. Using (2.6) with λ = 1 we obtain
kukL16
16 . kukX 5/16+,1/2+ . kukX s,1/2+ ,
t Lx
for s >
5
16 .
Note that we also use (2.3) in order to bound kJ s ukL4t L4x . Thus
k|u|4 ukX s,−1/2+2ǫ . kuk5X s,1/2+
which combined with (3.7) gives
1
kIukX 1,1/2+ . kIu0 kH 1 + δ 2 −ǫ kIuk5X 1,1/2+ .
Now by standard nonlinear techniques, see, for example, [12] we have that for δ ∼ 1
kIukX 1,1/2+ . kIu0 kH 1 .
Corollary 3.5. Consider the I-initial value problem (3.6). If kIu0 kH 1 . 1, then (3.6)
5
in [0, δ] ∼ [0, λ1aǫ ], where a is a fixed positive integer
is locally well-posed for any s > 16
(a ∼ 20 will do it).
Proof. The proof is line by line the same as in the previous proposition. The important step
is that in the iteration process where x ∈ [0, λ], δ has a fixed power and this fact gives us
1
for all practical
a local well-posedness result on the time interval [0, λ1aǫ ], where λ1aǫ ∼ λ0+
purposes.
10
DANIELA DE SILVA, NATAŠA PAVLOVIĆ, GIGLIOLA STAFFILANI, AND NIKOLAOS TZIRAKIS
Step 2: E2 (u) is a small perturbation of E1 (u). Now we shall prove that the second
energy E 2 (u) is a small perturbation of the first energy E 1 (u). This is the content of the
following proposition.
Proposition 3.6. Assume that u solves (3.6) with s ≥ 2/5. Then
E 2 (u) = E 1 (u) + O(1/N )kIuk6H 1 .
Moreover if kIukH 1 = O(1) then
1
).
N
Proof. By the definition of first and second modified energy we have
k∂x Iuk2L2 . E 2 (u) + O(
(3.8)
6
(3.9)
Y
1
1
1
E 2 (u) = − Λ2 (m1 k1 m2 k2 ) + Λ6 (M6 ) = E 1 (u) + Λ6 (M6 −
mi ).
2
6
6
i=1
Therefore it suffices to prove the following pointwise in time estimate
(3.10)
|Λ6 (M6 −
6
Y
mi )|(t) . O(1/N )kIu(·, t)k6Hx1 .
i=1
Combining the Decomposition Remark 2.1 with the fact that M6 is bounded (by Proposition 3.2) and that m is bounded (by its definition) it is enough to show that
Z
6
Y
Γ6 j=1
u
b(kj ) .
1
kIu(·, t)k6Hx1 .
N
Towards this aim, we break all the functions into a sum of dyadic constituents uj , each
with frequency support hki ∼ 2j , j = 0, .... We will be able to sum over all frequency pieces
uj of u since our estimate will decay geometrically in these frequencies. Hence we need to
show the following:
Z Y
6
6
Y
1
0−
ubj (kj ) . (N1 ...N6 )
(3.11)
kIuj (·, t)k6Hx1
N
Γ6
j=1
j=1
2hj
where uj is supported around hki ∼ Nj =
for some hj ’s.
⋆
⋆
⋆
⋆
⋆
⋆
Denote by N1 ≥ N2 ≥ N3 ≥ N4 ≥ N5 ≥ N6 the decreasing rearrangement of N1 , ..., N6 ,
and by u⋆j the function uj supported in Fourier space around Nj⋆ , j = 1, ..., 6.
Since we are integrating over Γ6 , we have that N1⋆ ∼ N2⋆ . Moreover, we may assume that
Q
N1⋆ ∼ N2⋆ & N , otherwise M6 − 6i=1 mi ≡ 0 and (3.10) follows trivially.
1
1
Since m(N ⋆ )(N
⋆ )1− . N 1− we have that
1
Z
1
6
Y
Γ6 j=1
(N1⋆ )0−
ubj (kj ) .
N 1−
Z
[⋆ )
(JIu
1
6
Y
j=2
6
⋆ 0−
Y
c⋆ . (N1 ) kIu⋆ kH 1
ku⋆j kL10
u
1
j
x
x
N 1−
j=2
by reversing Plancherel and applying Hölder’s inequality. Moreover by Sobolev embedding
ku⋆j kL10
. ku⋆j kH 2/5
x
x
and for s ≥ 2/5 we have
ku⋆j kH 2/5 . kIu⋆j k
x
Thus (3.11) follows.
1+ 2
5 −s
Hx
. kIu⋆j kHx1 .
GLOBAL WELL-POSEDNESS FOR A PERIODIC NONLINEAR SCHRÖDINGER EQUATION
11
Step 3: Decay of E2 (u). In order to estimate the decay of E 2 (u) in time we need to
bound the right hand side of (3.5). Towards this aim we prove a bilinear estimate using an
elementary number theory argument. Let η ∈ C0∞ be a bump function that is supported
on [−2, 2] and equals 1 on [−1, 1].
Proposition 3.7. Let φ1 and φ2 be λ−periodic functions whose Fourier transforms are
supported on {k : |k| ∼ N1 } and {k : |k| ∼ N2 } respectively with N1 >> N2 . Then
kη(t)(Uλ (t)φ1 ) η(t)(Uλ (t)φ2 )kL2t L2x . C(λ, N1 )kφ1 kL2 kφ2 kL2
where
C(λ, N1 ) =
(
1,
if N1 ≤ 1
1
1 12
( N1 + λ ) , if N1 > 1
)
.
Moreover,
kφ1 φ2 kL2t L2x . C(λ, N1 )kφ1 kX 0,1/2+ kφ2 kX 0,1/2+ .
Remark. We observe that as λ → ∞ we recover the refined bilinear estimate in R with
constant C(N1 ) = ( N11 )1/2 , see, for example, [4, 10, 23].
Proof. Let ψ be a positive even Schwartz function such that ψ = η̂. Then we have
B = kη(t)(Uλ (t)φ1 ) η(t)(Uλ (t)φ2 )kL2t L2x
Z
2
2
c
c
=
φ1 (k1 )φ2 (k2 )ψ(τ1 − k1 ) ψ(τ2 − k2 ) (dk1 )λ (dk2 )λ dτ1 dτ2 2
k=k1 +k2 , τ =τ1 +τ2
L2τ Lk
Z
1/2
2
2
e
×
ψ(τ − k1 − k2 ) (dk1 )λ (dk2 )λ
.
k=k1 +k2
(3.12)
×
1/2 e − k2 − k2 ) |φb1 (k1 )|2 |φb2 (k2 )|2 (dk1 )λ (dk2 )λ
ψ(τ
1
2
k=k1 +k2
Z
,
L2τ L2k
where to obtain (3.12) we used Cauchy-Schwartz and the following definition of ψe ∈ S
Z
e − k2 − k2 ).
ψ(τ1 − k12 ) ψ(τ2 − k22 ) dτ1 dτ2 = ψ(τ
1
2
τ =τ1 +τ2
An application of Hölder gives us the following upper bound on (3.12)
Z
1/2 e − k2 − k2 ) |φb1 (k1 )|2 |φb2 (k2 )|2 (dk1 )λ (dk2 )λ
ψ(τ
(3.13)
M
1
2
k=k1 +k2
,
L2τ L2k
where
Z
M =
k=k1 +k2
e −
ψ(τ
k12
− k22 )
1/2
(dk1 )λ (dk2 )λ ∞
.
Lτ L∞
k
Now by integration in τ followed by Fubini in k1 , k2 and two applications of Plancharel we
have
Z
1/2 e − k2 − k2 ) |φb1 (k1 )|2 |φb2 (k2 )|2 (dk1 )λ (dk2 )λ
ψ(τ
. kφ1 kL2x kφ2 kL2x ,
1
2
k=k1 +k2
2
Lk,τ
12
DANIELA DE SILVA, NATAŠA PAVLOVIĆ, GIGLIOLA STAFFILANI, AND NIKOLAOS TZIRAKIS
which combined with (3.12), (3.13) gives
(3.14)
B . M kφ1 kL2x kφ2 kL2x .
We find an upper bound on M as follows:
(3.15)
M.
1
sup #S
λ τ,k
!1/2
,
where
1
Z | |k1 | ∼ N1 , |k − k1 | ∼ N2 , k2 − 2k1 (k − k1 ) = τ + O(1)},
λ
and #S denotes the number of elements of S. When N1 ≤ 1 then #S . O(λ), which
implies C(λ, N1 ) . 1. When N1 > 1 rename k1 = z. Then
1
S = {z ∈ Z | |z| ∼ N1 , |k − z| ∼ N2 , k2 − 2z(k − z) = τ + O(1)}.
λ
Let z0 be an element of S i.e.
S = {k1 ∈
(3.16)
|z0 | ∼ N1 , |k − z0 | ∼ N2 ,
and
k2 + 2z02 − 2kz0 = τ + O(1).
(3.17)
In order to obtain an upper bound on #S, we shall count the number of z̄’s ∈
that z0 + z̄ ∈ S. Thus
(3.18)
|z0 + z̄| ∼ N1 , |z0 + z̄ − k| ∼ N2 ,
and
(3.19)
k2 + 2 (z0 + z̄)2 − 2k (z0 + z̄) = τ + O(1).
However by (3.17) we can rewrite the left hand side of (3.19) as follows
k2 + 2 (z0 + z̄)2 − 2k (z0 + z̄) = k2 + 2z02 + 2 (z̄)2 + 4z0 z̄ − 2kz0 − 2kz̄
= τ + O(1) + 2z̄ 2 + 4z0 z̄ − 2kz̄.
Hence it suffices to count z̄’s ∈ λ1 Z satisfying (3.18) and such that
k
z̄ 2 + 2z̄(z0 − ) = O(1),
2
(3.20)
where z0 satisfies (3.16) - (3.17).
By (3.16) and (3.18) we have that
|z̄| = |(z̄ + z0 − k) − z0 + k| . N2 + N2 .
Hence
(3.21)
|z̄| . N2 << N1 .
On the other hand, by (3.16) we have that
k
|z0 − | ∼ N1 .
2
(3.22)
Now rewrite (3.20) as follows
(3.23)
k
z̄ z̄ + 2(z0 − ) = O(1).
2
1
λ Z,
such
GLOBAL WELL-POSEDNESS FOR A PERIODIC NONLINEAR SCHRÖDINGER EQUATION
13
Therefore, using (3.21)-(3.22), the estimate (3.23) implies that
|z̄| N1 = O(1).
Since z̄ ∈
1
λZ
this implies that the number of z̄’s of size 1/N1 is λ/N1 . Hence
#S . 1 +
λ
,
N1
which combined with (3.15) gives
M≤
and therefore
1
1
+
N1 λ
C(λ, N1 ) ≤
1/2
1
1
+
N1 λ
,
1
2
.
Now we are ready to prove desired decay of E 2 (u) which follows from the following
proposition:
Proposition 3.8. For any λ-periodic Schwartz function u, with period λ ≥ N , and any δ
given by the local theory, we have that
Z δ
,
Λ10 (M10 ; u(t))| . λ0+ N −5/2+ kIuk10
|
X 1,1/2+
0
for s > 11/28.
Proof. We perform a dyadic decomposition as in Proposition 3.6, and we borrow the same
notation. From now on we do not keep track of all the different λkǫ , k ∈ R, and we just
write λ0+
We observe that M10 is bounded as an elongation of the bounded multiplier M6 . Since
the multiplier M10 vanishes on Γ10 when all frequencies are smaller than N , we can assume
that there are N1∗ , N2∗ & N . Now we divide the proof in two cases.
Case 1. N1⋆ ∼ N2⋆ ∼ N3⋆ & N
Since
1
m(N1⋆ )m(N2⋆ )m(N3⋆ )(N1⋆ N2⋆ N3⋆ )1−
|
Z
δ
0
Z
M10
10
Y
j=1
.
1
N 3−
we have
10
Y
(N1⋆ )0−
⋆
⋆
⋆
ûj | .
ku⋆j kL14
kJIu
k
6 L6 kJIu2 kL6 L6 kJIu3 kL6 L6
14
1
L
t Lx
t x
t x
t x
N 3−
j=4
(N ⋆ )0−
λ0+ 13− kIuk3X 1,1/2+ kuk7 ( 1 − 3 )+,1/2+
X 2 14
N
(3.24)
.
(3.25)
. λ0+ (N1⋆ )0− N −3+ kIuk10
X 1,1/2+
where in order to obtain (3.24) we use (2.6) and to obtain (3.25) we use the fact that for
s > 2/7 the following inequality holds
kuk
1
3
X ( 2 − 14 )+,1/2+
Case 2. N1⋆ ∼ N2⋆ ≫ N3⋆
= kukX 2/7+,1/2+ . kIukX 1,1/2+ .
14
DANIELA DE SILVA, NATAŠA PAVLOVIĆ, GIGLIOLA STAFFILANI, AND NIKOLAOS TZIRAKIS
Since δ < 1 we can insert η(t) where η is a bump function supported in [−1/2, 3/2] and
equals 1 in [0, 1]. Also notice that m(N ⋆ )m(N 1⋆ )(N ⋆ N ⋆ )1− . N12− . Hence using the Cauchy1
2
1 2
Schwartz inequality we get:
Z
Z δ
10
10
Y
Y
(N1⋆ )0−
⋆ ⋆
⋆
ûj | .
η(t) M10
|
u⋆i kL2t L2x
kηJIu
u
k
2
kJIu
1 3 Lt L2x
2
2−
N
0
4
j=1
10
(3.26)
.
(3.27)
.
(3.28)
.
Y
(N1⋆ )0− 1
1 1
⋆
⋆
⋆
2 kJIu k 0,1/2+ ku k 0,1/2+ kJIu
)
(
+
u⋆i kL2t L2x
2
3
1
X
X
∗
N 2− λ N1
4
10
0−
⋆
Y
(N1 )
ku⋆j kL28
kIu⋆1 kX 1,1/2+ ku⋆3 kX 0,1/2+ kJIu⋆2 kL4t L4x
28
t Lx
N 5/2−
j=4
10
⋆ 0−
Y
0+ (N1 )
3
λ
kuj kX 1/2−3/28,1/2+
kIukX 1,1/2+
N 5/2−
j=4
. λ0+ (N1⋆ )0− N −5/2+ kIuk10
X 1,1/2+
where to obtain (3.26) we use Proposition 3.7, to obtain (3.27) we use Hölder’s inequality
and the facts that λ ≥ N and N1∗ & N , to obtain (3.28) we use (2.6), and in the last line
we use that s > 11/28.
Now we are ready to prove our main theorem on T1 .
Step 4: Proof of Theorem 1.
Proof. Let u0 ∈ H s , where 4/9 < s < 1/2. By definition,
|m(k)|
|k|s−1
. N 1−s , hence
1−s
cλ k ≤ N 1−s k|k|s u
cλ k . N
cλ k = k |m(k)| |k|s u
ku0 kḢ s .
k∂x Iuλ0 k2 = km(k)|k|u
0 2
0 2
0 2
s−1
|k|
λs
By the Gagliardo-Nirenberg inequality we have:
N 2−2s
ku0 k2Ḣ s ku0 k42 .
λ2s
Combining the two inequalities above with the definition of the first modified energy we
obtain
Z
Z
1
1
N 2−2s
1 λ
λ 2
E (u0 ) =
|∂x Iu0 | dx +
|Iuλ0 |6 dx .
ku0 k2Ḣ s .
2 R
6 R
λ2s
kIuλ0 k66 . k∂x Iuλ0 k22 kIuλ0 k42 .
We choose λ ∼ N
Therefore
(3.29)
1−s
s
, which for s < 1/2 implies λ & N . Then we have that E 1 (uλ0 ) . 1.
kIuλ0 k2H 1 . 1,
which allows us to apply Propositions 3.4, 3.6. The main idea of the proof is that if in each
step of the iteration we have the same bound for kIuλ0 kH 1 then we can iterate again with
the same timestep. Because we iterate the uλ solutions, the interval of local existence is of
order λ1aǫ . But this creates no additional problem since λ is given in terms of N . Moreover
N is a fixed number and thus the interval of local existence does not shrink. This is the
reason why we don’t write down the λ0+ dependence in the following string of inequalities.
Now, by Lemma 3.1, Proposition 3.6, Proposition 3.8 and (3.29) we have that
1
E 2 (uλ (δ)) . E 2 (uλ (0)) + CN −5/2+ . E 1 (uλ (0)) + kIuλ0 k6H 1 + CN −5/2+ . 1,
N
GLOBAL WELL-POSEDNESS FOR A PERIODIC NONLINEAR SCHRÖDINGER EQUATION
15
for N large enough. Combining the inequality above with Proposition 3.6 we have,
1
k∂x Iu(δ)λ k22 . E 2 (uλ (δ)) + O( ) . 1.
N
Thus
kIuλ (δ)kH 1 . 1
and we can continue the solution in [0, M δ] = [0, T ] as long as T ≪ N 5/2− . Hence
kIuλ (T )kH 1 . 1
for all T ≪ N 5/2− . From the definition of I this implies that
kuλ (T )kH s . 1
for all T ≪ N 5/2− . Undoing the scaling we have that
ku(T )kH s . CN,λ
for all T ≪
N 5/2−
λ2 .
But
N 5/2−
λ2
∼N
9s−4
2s
goes to infinity as N → ∞ since s > 4/9.
4. The I-method and the proof of Theorem 1.2.
In this section we present the proof of Theorem 1.2. We will focus on the analysis of the
λ-periodic problem:
(4.1)
iut + ∆u − |u|2 u = 0
(4.2)
u(x, 0) = u0 (x) ∈ H s (T2λ ), t ∈ R,
for 0 < s < 1. As for the one dimensional case, we will use the I-method. The definitions of
the multiplier m and of the operator I are the same as in the 1D context. Precisely, given
a parameter N >> 1 to be chosen later, we define m(k) to be the following multiplier:
m(k) =
(
1
|k| s−1
N
if |k| < N
if |k| > 2N
Then I : H s → H 1 will be defined as the following multiplier operator:
c
(Iu)(k)
= m(k)û(k).
This operator is smoothing of order 1 − s, that is:
(4.3)
kuks0 ,b0 . kIuks0 +1−s,b0 . N 1−s kuks0 ,b0
for any s0 , b0 ∈ R.
The first modified energy of a function u ∈ H s is defined by
E 1 (u) = E(Iu),
where we recall that
Z
Z
1
1
2
E(u)(t) =
|∇u(t)| dx +
|u(t)|4 dx.
2
4
Our main objective will be to prove that the modified energy of a solution to the λperiodic problem (4.1)-(4.2) is almost conserved in time.
We now proceed to present the steps leading to the proof of Theorem 1.2.
16
DANIELA DE SILVA, NATAŠA PAVLOVIĆ, GIGLIOLA STAFFILANI, AND NIKOLAOS TZIRAKIS
Step 1: Local well-posedness for the I-system. The first step towards the proof of
Theorem 1.2 is to apply the I-operator to (1.1) and prove a local well-posedness result for
the I-initial value problem
iIut + I∆u − I(|u|2 u) = 0
(4.4)
Iu(x, 0) = Iu0 (x) ∈ H s (T2λ ), t ∈ R.
This is the content of the next proposition.
Proposition 4.1. Consider the I-initial value problem (4.4), λ=1. If kIu0 kH 1 . 1, then
7
in [0, δ] ∼ [0, 1].
(4.4) is locally well-posed for any s > 20
Proof. By Duhamel’s formula as in Proposition 3.4 we have that
1
kIukX 1,1/2+ . kIu0 kH 1 + δ 20 −ǫ kI(|u|2 u)k
9
X 1,− 20
By the “invariant lemma” in [11] we know that the estimate
kI(|u|2 u)k
9
X 1,− 20
. kIuk3X 1,1/2+
is implied by
k|u|2 uk
(4.5)
9
X s,− 20
. kuk3X s,1/2+ .
Since the rest of the proof is identical to Proposition 3.4 we briefly recall how (4.5) can be
obtained. Note that this estimate was first proved by Bourgain [2] and his proof is based
upon a local variant of the well known periodic Strichartz estimate

1
2
XZ
2 2b1
2
s1 
dτ (1 + |τ − |k| |) |û(k, τ )|
(4.6)
kukL4t L4x . N
k∈Q
where b1 >
1−min(1/2,s1 )
.
2
Here we omit the proof and refer the reader to [2] for details.
Corollary 4.2. Consider the I-initial value problem (4.4). If kIu0 kH 1 . 1, then (4.4)
7
is locally well-posed for any s > 20
in [0, δ] ∼ [0, λ1aǫ ], where a is a fixed positive integer
(a ∼ 100 will do it).
Proof. The proof is identical to the argument of Bourgain in [2], the only difference being
the fact that (4.6) holds true for the λ-periodic solutions, but now with a factor of order
λ0+ on the right hand side. Again the fixed power of δ in the local theory, enables us to
prove the local well-posedness result for the λ-periodic problem in the interval [0, λ1aǫ ]. Step 2: Decay of the first energy. In order to prove that the first energy is almost
conserved, we will use the bilinear estimate for λ-periodic functions stated in Proposition
4.6. Its proof is based on some number theoretic facts that we recall in the following three
lemmas; see also related estimates in the work of Bourgain [6].
The following lemma is known as Pick’s Lemma [24]:
Lemma 4.3. Let Ar be the area of a simply connected lattice polygon. Let E denote the
number of lattice points on the polygon edges and I the number of lattice points in the
interior of the polygon. Then
1
Ar = I + E − 1.
2
1/3
,
Lemma 4.4. Let C be a circle of radius R. If γ is an arc on C of length |γ| < 34 R
then γ contains at most 2 lattice points.
GLOBAL WELL-POSEDNESS FOR A PERIODIC NONLINEAR SCHRÖDINGER EQUATION
17
γ
A
B
θ θ
O
C
Figure 1. Triangle area.
Proof. We prove the lemma by contradiction. Assume that there are 3 lattice points P1 ,
P2 and P3 on an arc γ = AB of C, and denote by T (P1 , P2 , P3 ) the triangle with vertices
P1 , P2 and P3 . Then, by Lemma 4.3 we have
1
3
1
1
Area of T (P1 , P2 , P3 ) = I + E − 1 ≥ I + − 1 = I + ≥ .
2
2
2
2
1/3
, then
We shall prove that under the assumption that |γ| < 43 R
1
Area of T (P1 , P2 , P3 ) < ,
2
hence γ must contain at most two lattice points.
We observe that (see Figure 1)
(4.7)
Area of the sector ABO = R2 θ,
Area of the triangle ABO = R2 sin θ cos θ.
Hence, for any P1 , P2 , P3 on γ we have
(4.8)
Area of T (P1 , P2 , P3 ) ≤ R2 θ − R2 sin θ cos θ = R2 (θ −
1
sin(2θ)).
2
One can easily check that
1
sin(2θ) ≤
2
Thus (4.8), (4.9) and the fact that |γ| = Rθ imply
(4.9)
θ−
2 3
θ .
3
that
2
1
2
Area of T (P1 , P2 , P3 ) ≤ R2 θ 3 = R2 (|γ|R−1 )3 < ,
3
3
2
where to obtain the last inequality we used the assumption that |γ| <
(4.7) is proved.
Also we recall the following result of Gauss, see, for example [19]
Lemma 4.5. Let K be a convex domain in R2 . If
N (λ) = #{Z2 ∩ λK},
1/3
3
.
4R
Therefore
18
DANIELA DE SILVA, NATAŠA PAVLOVIĆ, GIGLIOLA STAFFILANI, AND NIKOLAOS TZIRAKIS
then, for λ >> 1
N (λ) = λ2 |K| + O(λ),
where |K| denotes the area of K and #S denotes the number of points of a set S.
Now we are ready to present the bilinear estimate for λ-periodic functions. Let η ∈ C0∞
be a bump function that is supported on [−2, 2] and equals 1 on [−1, 1].
Proposition 4.6. Let φ1 and φ2 be λ−periodic functions, λ ≥ 1, whose Fourier transforms
are supported on {k : |k| ∼ N1 } and {k : |k| ∼ N2 } respectively, with N1 >> N2 > 1.
(a) Then
kη(t)(Uλ (t)φ1 ) η(t)(Uλ (t)φ2 )kL2t L2x . (λN2 )ǫ kφ1 kL2 kφ2 kL2 ,
(4.10)
for any ǫ > 0. Hence
kφ1 φ2 kL2t L2x . (λN2 )ǫ kφ1 kX 0,1/2+ kφ2 kX 0,1/2+ .
(b) Moreover, if λ >> 1 then
(4.11)
kη(t)(Uλ (t)φ1 ) η(t)(Uλ (t)φ2 )kL2t L2x .
1 N2
+
λ N1
1/2
kφ1 kL2 kφ2 kL2 .
Remark. Note that as λ → ∞ we recover the improved bilinear Strichartz with constant
2 1/2
C(N1 , N2 ) = ( N
.
N1 )
Proof. (a) Let ψ be a positive even Schwartz function such that ψ = η̂. Then we have
B = kη(t)(Uλ (t)φ1 ) η(t)(Uλ (t)φ2 )kL2t L2x
Z
c1 (k1 )φ
c2 (k2 )ψ(τ1 − k2 ) ψ(τ2 − k2 ) (dk1 )λ (dk2 )λ dτ1 dτ2 k 2 2
φ
=k
1
2
Lτ L
k
k=k1 +k2 , τ =τ1 +τ2
.k
Z
k=k1 +k2
(4.12)
×
Z
k=k1 +k2
e − k2 − k2 ) (dk1 )λ (dk2 )λ
ψ(τ
1
2
1/2
×
e − k2 − k2 ) |φb1 (k1 )|2 |φb2 (k2 )|2 (dk1 )λ (dk2 )λ
ψ(τ
1
2
1/2
kL2τ L2
k
where to obtain (4.12) we used Cauchy-Schwartz and the following definition of ψe ∈ S
Z
e − k2 − k2 ).
ψ(τ1 − k12 ) ψ(τ2 − k22 ) dτ1 dτ2 = ψ(τ
2
1
τ =τ1 +τ2
An application of Hölder gives us the following upper bound on (4.12)
1/2
Z
2
2 b
2 b
2
e
kL2τ L2 ,
ψ(τ − k1 − k2 ) |φ1 (k1 )| |φ2 (k2 )| (dk1 )λ (dk2 )λ
(4.13)
Mk
k
k=k1 +k2
where
M =k
Z
k=k1 +k2
e − k2 − k2 ) (dk1 )λ (dk2 )λ k1/2
ψ(τ
∞.
1
2
L∞
τ L
k
Now by integration in τ followed by Fubini in k1 , k2 and two applications of Plancharel we
have
1/2
Z
2
2 b
2 b
2
e
kL2τ L2 . kφ1 kL2x kφ2 kL2x ,
ψ(τ − k1 − k2 ) |φ1 (k1 )| |φ2 (k2 )| (dk1 )λ (dk2 )λ
k
k=k1 +k2
k
GLOBAL WELL-POSEDNESS FOR A PERIODIC NONLINEAR SCHRÖDINGER EQUATION
19
which combined with (4.12) and (4.13) gives
(4.14)
B . M kφ1 kL2x kφ2 kL2x .
We find an upper bound on M as follows:
(4.15)
M.
1
sup #S
λ2 τ,k
!1
2
,
where
S = {k1 ∈ Z2 /λ | |k1 | ∼ N1 , |k − k1 | ∼ N2 , |k|2 − 2k1 · (k − k1 ) = τ + O(1)},
and #A denotes the number of lattice points of a set A.
For notational purposes, let us rename k1 = z, that is
S = {z ∈ Z2 /λ | |z| ∼ N1 , |k − z| ∼ N2 , |k|2 + 2|z|2 − 2k · z = τ + O(1)}.
Let z0 be an element of S i.e.
(4.16)
|z0 | ∼ N1 , |k − z0 | ∼ N2 ,
and
|k|2 + 2|z0 |2 − 2k · z0 = τ + O(1).
(4.17)
In order to obtain an upper bound on #S, we shall count the number of l’s ∈ Z2 such that
z0 + λl ∈ S where z0 satisfies (4.16) - (4.17). Thus such l’s must satisfy
l
l (4.18)
z0 + λ ∼ N1 , z0 + λ − k ∼ N2 ,
and
(4.19)
l
l 2
|k| + 2 z0 + − 2k · (z0 + ) = τ + O(1).
λ
λ
2
However by (4.17) we can rewrite the left hand side of (4.19) as follows
2
l
l
l
l 2
l
2
2
2
|k| + 2 z0 + − 2k · (z0 + ) = |k| + 2|z0 | + 2 + 4z0 · − 2k · z0 − 2k ·
λ
λ
λ
λ
λ
2
l
l
l
= τ + O(1) + 2 + 4z0 · − 2k · .
λ
λ
λ
Therefore (4.19) holds if
(4.20)
2
l
+ 2 l · (z0 − k ) = O(1).
λ
λ
2
Moreover, (4.16) and (4.18) yield
l l
= + z0 − k − z0 + k . N2 + N2 ,
λ λ
that is
(4.21)
|l| . λN2 .
Finally we observe that (4.16) together with the assumption that N1 >> N2 implies that
z0 k z0 k z0 k z0 N1 ∼ N1 − N2 ∼ − − ≤ z0 − ≤ − + ∼ N2 + N1 . N1 ,
2
2
2
2
2
2
2
20
DANIELA DE SILVA, NATAŠA PAVLOVIĆ, GIGLIOLA STAFFILANI, AND NIKOLAOS TZIRAKIS
i.e.
z0 −
(4.22)
k ∼ N1 .
2
Hence, it suffices to count the l′ s ∈ Z2 satisfying (4.20) and (4.21) where z0 is such that
(4.22) holds.
Let w = (a, b) denote the vector z0 − k2 . Thus we need to count the number of points in
the set Aλ
(4.23)
Aλ = {l ∈ Z2 : |l|2 + 2λl · w = O(λ2 ), |l| . λN2 , |w| ∼ N1 },
or equivalently,
(4.24)
Aλ = {(x, y) ∈ Z2 : x2 + y 2 + 2λ(ax + by) ≤ cλ2 , x2 + y 2 ≤ (k2 λN2 )2 , a2 + b2 ∼ N12 },
λ , C λ be the following circles,
for some c, k2 > 0. Let C−
+
λ
C−
: (x + λa)2 + (y + λb)2 = −cλ2 + λ2 (a2 + b2 )
λ
: (x + λa)2 + (y + λb)2 = cλ2 + λ2 (a2 + b2 )
C+
and for any integer n, let Cnλ be the circle
Cnλ : (x + λa)2 + (y + λb)2 = n + λ2 (a2 + b2 ).
Finally, let D λ denote the disk
D λ : x2 + y 2 ≤ (k2 λN2 )2 .
B
γn
Dλ
λ
C+
D
Cnλ
θM
C0λ
C
λ
C−
Figure 2. Circular sector
GLOBAL WELL-POSEDNESS FOR A PERIODIC NONLINEAR SCHRÖDINGER EQUATION
21
We need to count the number of lattice points inside D λ that are on arcs of circles Cnλ ,
with
−cλ2 ≤ n ≤ cλ2 .
Precisely, the total number of lattice point in Aλ can be bounded from above by
2cλ2 × #(Cnλ ∩ D λ ).
(4.25)
Denote by γn the arc of circle Cnλ which is contained in D λ . Notice that (see Figure 2)
(4.26)
p
|γn | ≤ RM θM
where RM = cλ2 + k1 λ2 N12 for some constant k1 > 0, and θM is the angle between the
line segment CB and CD, which lie along the tangent lines from C = (−λa, −λb) to the
circle x2 + y 2 = (k2 λN2 )2 . Hence,
sin θM ≤ k
N2
,
N1
for some constant k > 0. Since N1 >> N2 , we can assume that sin θM > 12 θM . Hence,
(4.27)
θM < 2k
N2
.
N1
In order to count efficiently the number of lattice points on each γn , we distinguish two
cases based on the application of Lemma 4.4.
2
1
3 3 −3
2
Case 1: 2k N
RM .
N1 < 4
In this case (4.26)-(4.27) guarantee that the hypothesis of Lemma 4.4 is satisfied by each
arc of circle γn . Hence, on each γn there are at most two lattice points.
2
1
3 3 −3
2
Case 2: 2k N
≥
R
M .
N1
4
In this case we approximate the number of lattice points on γn by the number of lattice
points on Cnλ (see for example [1] , [7]):
(4.28)
ǫ
#Cnλ . RM
∼ (λN1 )ǫ . (λN2 )3ǫ
for any ǫ > 0.
Combining the estimate in (4.25), Case 1 and Case 2 we conclude that
#S . λ2 + λ2 (λN2 )ǫ ,
for any ǫ > 0. Since λ, N2 ≥ 1, together with (4.15), this implies that
M . (λN2 )ǫ ,
for all positive ǫ’s. Hence (4.10) follows.
(b) As in part (a) we will count the number of points in the set Aλ given by (4.23).
According to (4.24) we have that
(4.29)
with
Aλ = Z2 ∩ λB1 ,
(4.30) B1 = {(x, y) ∈ R2 : x2 + y 2 + 2(ax + by) ≤ c, x2 + y 2 ≤ (k2 N2 )2 , a2 + b2 ∼ N12 }.
22
DANIELA DE SILVA, NATAŠA PAVLOVIĆ, GIGLIOLA STAFFILANI, AND NIKOLAOS TZIRAKIS
In the rest of the proof for arbitrary two points P1 , P2 on a circle with center C we
denote by S(C, P1 , P2 ) the solid sector contained between the lines CP1 ,CP2 and the circle
arc between P1 and P2 .
B
1
C+
A
D
C01
E
1
C−
C
Figure 3. Circular sectors when λ = 1
We observe that (see Figure 3 )
B1 ⊂ S(C, B, D) \ S(C, A, E),
where CB and CD lie along the tangent lines from C = (−a, −b) to the circle x2 + y 2 =
(k2 N2 )2 . Hence, setting S1 = S(C, B, D), S2 = S(C, A, E), and H = S1 \ S2 we have
#Aλ = #{Z2 ∩ λB1 }
≤ #{Z2 ∩ λS1 } − #{Z2 ∩ λS2 }
(4.31)
= λ2 |S1 | − λ2 |S2 | + O(λ)
(4.32)
= λ2 |H| + O(λ),
where to obtain (4.31) we used Lemma 4.5 twice. In order to compute the area of H notice
that since a2 + b2 ∼ N12 , we have (see Figure 3)
N2
sin θ ∼
.
N1
Hence, by the assumption N1 >> N2 , we get θ ∼
N2
N1 .
|H| = (r12 − r22 )θ ∼
with
Therefore
N2
,
N1
r12 = c + (a2 + b2 ), r22 = −c + (a2 + b2 ).
GLOBAL WELL-POSEDNESS FOR A PERIODIC NONLINEAR SCHRÖDINGER EQUATION
23
This combined with (4.32) implies that
#Aλ . λ + λ2
N2
,
N1
that together with (4.15) gives
M.
1 N2
+
λ N1
1/2
.
Now we are ready to prove the desired decay of the first modified energy, which is the
content of the following proposition:
Proposition 4.7. Let s > 1/2, λ . N, t > 0, and u0 ∈ H s (T2λ ) be given. If u is a solution
to (4.1) − (4.2) then
1
|E 1 (u)(t) − E 1 (u)(0)| . 1− kIukX 1,1/2+ .
N
Proof. The definition of E together with equation (4.1) give that:
Z
I(u)t (|Iu|2 Iu − ∆Iu − iIut )
∂t E(Iu)(t) = Re
= Re
Z
T2λ
T2λ
I(u)t (|Iu|2 Iu − I(|u|2 u))
Therefore, integrating in time, and using the Parseval’s formula we get:
(4.33)
E 1 (u)(t) − E 1 (u)(0) =
Z tZ m(k2 + k3 + k4 )
du(k )Iu(k
c )Iu(k
c 2 )Iu(k
c 4)
I∂
=
1−
t
1
3
m(k
)m(k
)m(k
)
2
3
4
0
Γ4
Using the equation (4.1), we then get:
E 1 (u)(t) − E 1 (u)(0) = Tr1 + Tr2 ,
where
(4.34)
Tr1 =
Z tZ
=
0
(4.35)
Γ4
Tr2 =
Z tZ
=
0
Γ4
We wish to prove that:
m(k2 + k3 + k4 )
1−
m(k2 )m(k3 )m(k4 )
m(k2 + k3 + k4 )
1−
m(k2 )m(k3 )m(k4 )
d
c
c
c
∆Iu(k
1 )Iu(k2 )Iu(k3 )Iu(k4 )
\
c )Iu(k
2 u)(k )Iu(k
c 2 )Iu(k
c 4)
I(|u|
1
3
1
,
N 1−
for a constant C = C(kIukX 1,1/2+ ). According to the Decomposition Remark in our
estimates we may ignore the presence of the complex conjugates.
In order to obtain the estimate (4.36), we break u into a sum of dyadic pieces uj , each with
frequency support hki ∼ 2j , j = 0, .... We will then be able to sum over all the frequency
(4.36)
|Tr1 | + |Tr2 | .
24
DANIELA DE SILVA, NATAŠA PAVLOVIĆ, GIGLIOLA STAFFILANI, AND NIKOLAOS TZIRAKIS
pieces uj of u, since our estimates will decay geometrically in these frequencies.
We start by analyzing Tr1 . First notice that,
k∆(Iu)kX −1,1/2+ ≤ kIukX 1,1/2+ ,
therefore, the estimate on Tr1 will follow, if we prove that:
Z t Z m(k
+
k
+
k
)
2
3
4
c
c
c
c
φ1 (k1 )φ2 (k2 )φ3 (k3 )φ4 (k4 )
(4.37)
1−
m(k2 )m(k3 )m(k4 )
0
.
Γ4
i=4
Y
1
0−
kφi kX 1,1/2+
(N
N
N
N
)
kφ
k
+
1
2
3
4
1
−1,1/2
X
N 1−
i=2
for any λ-periodic function φi , i = 1, ..., 4 with positive spatial Fourier transforms supported
on
hki ∼ 2li ≡ Ni ,
(4.38)
for some li ∈ {0, 1, ...}.
By the symmetry of the multiplier, we can assume that
(4.39)
N2 ≥ N3 ≥ N4 .
Moreover, since we are integrating over Γ4 , we have that N1 . N2 . Therefore, we only need
to get the factor N20− in the estimate (4.37), in order to be able to sum over all dyadic
pieces.
Now we consider three different cases, by comparing N to the size of the Ni ’s.
Case I. N >> N2 . In this case, the symbol in Tr1 is identically zero, and the desired
bound follows trivially.
Case II. N2 & N >> N3 ≥ N4 . Since we are on Γ4 , we also have that N1 ∼ N2 . By the
mean value theorem, we have
m(k2 ) − m(k2 + k3 + k4 ) m′ (k2 ) m(k
+
k
+
k
)
2
3
4
=
.
N3 ∼ N3 .
1 −
m(k2 )
m(k2 )
m(k2 ) N2
Using the pointwise bound above, the Cauchy-Schwartz inequality, and Plancharel’s theorem we get:
N3
kφ1 φ3 kL2t L2x kφ2 φ4 kL2t L2x
(4.40)
|Tr1 | .
N2
By two applications of Hölder’s inequality, followed by four applications of the Strichartz
estimate (2.8) we obtain:
|Tr1 | .
N3 0+ (N1 )1+
λ
kφ1 kX −1,1/2+ kφ3 kX 1,1/2+ kφ2 kX 1,1/2+ kφ4 kX 1,1/2+ .
N2
(N2 N3 N4 )1−
Since N1 ∼ N2 & N , and λ . N , it follows immediately that in this case
1
|Tr1 | . 1− N20− kφ1 kX −1,1/2+ kφ2 kX 1,1/2+ kφ3 kX 1,1/2+ kφ4 kX 1,1/2+ .
N
Case III. N2 ≥ N3 & N. We will use the following crude bound on the multiplier:
m(k1 )
m(k2 + k3 + k4 ) (4.41)
1 − m(k2 )m(k3 )m(k4 ) . m(k2 )m(k3 )m(k4 ) ,
GLOBAL WELL-POSEDNESS FOR A PERIODIC NONLINEAR SCHRÖDINGER EQUATION
25
which follows from the fact that we are integrating on Γ4 . We also need the following
estimate:
1
. N −α ,
(4.42)
m(k)|k|α
for α ≥ 1/2−, |k| & N .
We distinguish two subcases:
IIIa)N2 ∼ N3 & N. Using Plancharel’s, followed by Holder’s inequality, and then by four
applications of the Strichartz estimate (2.8), we get:
(N1 )1+
m(k1 )
λ0+
×
m(k2 )m(k3 )m(k4 )
(N2 N3 N4 )1−
× kφ1 kX −1,1/2+ kφ2 kX 1,1/2+ kφ3 kX 1,1/2+ kφ4 kX 1,1/2+
1
kφ1 kX −1,1/2+ kφ2 kX 1,1/2+ kφ3 kX 1,1/2+ kφ4 kX 1,1/2+ ,
. λ0+ 1−
N m(k4 )N41−
|Tr1 | .
(4.43)
where to obtain (4.43) we use that N1 ≤ N2 together with (4.42) with α = 1− (N3 & N ).
Now if N4 & N , we conclude with an application of the monotonicity property (4.42). If
N4 << N , then we use that m(k4 ) = 1 and N4 ≥ 1. Hence, since λ . N , we get
|Tr1 | .
1
N 0− kφ1 kX −1,1/2+ kφ2 kX 1,1/2+ kφ3 kX 1,1/2+ kφ4 kX 1,1/2+
N 1− 2
IIIb)N2 >> N3 & N . Since we are integrating on Γ4 , we get N1 ∼ N2 . Let us emphasize
that in the case IIIb) we could not use the same strategy as in the case IIIa). More
precisely, the approach of IIIa) consisting of Plancharel’s, followed by Holder’s inequality,
and then by four applications of the Strichartz estimate (2.8) would produce a bound of
the type
(N1 )1+
m(k1 )
λ0+
× kφ1 kX −1,1/2+ kφ2 kX 1,1/2+ kφ3 kX 1,1/2+ kφ4 kX 1,1/2+
m(k2 )m(k3 )m(k4 )
(N2 N3 N4 )1−
1
. 1− λ0+ N10+ kφ1 kX −1,1/2+ kφ2 kX 1,1/2+ kφ3 kX 1,1/2+ kφ4 kX 1,1/2+
N
which we cannot sum in N1 (here we cannot cancel N10+ by N30+ since N3 << N1 ). Thus
to handle the case IIIb) we do need an improved Strichartz estimate given in Proposition
4.6.
In order to obtain the desired estimate, we apply the Cauchy-Schwartz inequality, together with Proposition 4.6 for the pair φ1 , φ3 and for the pair φ2 , φ4 . Thus
m(k1 )
N1
(λN3 )ǫ (λN4 )ǫ
×
m(k2 )m(k3 )m(k4 )
N2 N3 N4
× kφ1 kX −1,1/2+ kφ2 kX 1,1/2+ kφ3 kX 1,1/2+ kφ4 kX 1,1/2+ .
1
1
.
λ0+ 1− 1− ×
m(k3 )m(k4 )
N3 N4
× kφ1 kX −1,1/2+ kφ2 kX 1,1/2+ kφ3 kX 1,1/2+ kφ4 kX 1,1/2+ .
1
1
. 1− λ0+
kφ1 kX −1,1/2+ kφ2 kX 1,1/2+ kφ3 kX 1,1/2+ kφ4 kX 1,1/2+
N
m(k4 )N41−
|Tr1 | .
(4.44)
(4.45)
26
DANIELA DE SILVA, NATAŠA PAVLOVIĆ, GIGLIOLA STAFFILANI, AND NIKOLAOS TZIRAKIS
where in order to obtain (4.44) we have used that N1 ∼ N2 and in order to obtain (4.45)
we have used (4.42) once with α = 1− (N3 & N ). Finally as observed in the case IIIa),
m(k4 )N41− ≥ 1. Therefore
1
N 0− kφ1 kX −1,1/2+ kφ2 kX 1,1/2+ kφ3 kX 1,1/2+ kφ4 kX 1,1/2+ .
N 1− 2
Now, we proceed to analyze Tr2 . We claim that the following stronger estimate holds:
1
|Tr2 | . 2− kIuk6X 1,1/2+ .
N
As for Tr1 it suffices to prove the following bound:
Z tZ
6
6
Y
N 0− Y
φbj . max
LHS =
m123 (m456 − m4 m5 m6 )
kIφi kX 1,1/2+ ,
N 2−
0
Γ6
|Tr1 | .
j=1
i=1
with the φi ’s having positive spatial Fourier transform, supported on
hki ∼ 2li ≡ Ni ,
for some li ∈ {0, 1, ...}, and with Nmax , Nmed denoting respectively the biggest and the
second biggest frequency among the Ni ’s.
Since we are integrating over Γ6 , we can assume that Nmax ∼ Nmed . Also, we only need to
consider the case when Nmax & N , otherwise the symbol in (LHS) is zero, and the estimate
above holds trivially.
Now, using Holder’s inequality, together with the estimate (4.42) twice, and the fact that
the multiplier m is bounded, we then obtain:
Z tZ Y
6
1− N 1− m
Nmax
med max mmed
φbj (k, t)
LHS ≤ 1− 1−
Nmax Nmed mmax mmed 0 Γ6 j=1
4
0−
Y
Nmax
1−
1−
kJ Iφmax kL4t L4x kJ Iφmed kL4t L4x
kφk kL8t L8x .
.
N 2−
k=1
The Strichartz estimate (2.8), and the interpolation estimate (2.10) then imply:
LHS .
4
6
0−
0− Y
Y
Nmax
Nmax
kIφj kX 1,1/2+ ,
kIφ
k
kIφ
k
.
kφ
k
max X 1,1/2+
med X 1,1/2+
k X 1/2+,1/2+
N 2−
N 2−
k=1
where in the last inequality we have used (4.3).
j=1
Step 3: Proof of Theorem 1.2. We are now ready to exhibit the proof of Theorem 1.2.
As we explained in Theorem 1.1 we do not keep track of the λ0+ factors.
Proof. Let u0 be an initial data, and let N >> 1 be given. Recall that, for any λ ∈ R, we
set uλ0 (x, t) = λ1 u0 ( λx , λt2 ). Now, let us choose a rescaling parameter λ so that E 1 (uλ0 ) . 1,
that is
1−s
λ∼N s .
Then, by Corollary 4.2 and Proposition 4.7, there exists a δ, and a solution uλ to (4.1),
with initial condition uλ0 , such that
1
1
1
.
E (uλ )(δ) . E (uλ )(0) + O
N 1−
GLOBAL WELL-POSEDNESS FOR A PERIODIC NONLINEAR SCHRÖDINGER EQUATION
27
Therefore, we can continue our solution uλ until the size of E 1 (uλ )(t) reaches 1, that is at
least C · N 1− times. Hence
E 1 (uλ )(C · N 1− δ) ∼ 1.
Given T >> 1, we choose our parameter N >> 1, so that
3s−2
C · N 1− δ
∼ N s −.
2
λ
We observe that the exponent of N is positive as long as s > 2/3, hence N is well-defined
for all times T. Now, undoing the scaling, we get that:
T ∼
E 1 (u)(T ) . T
4(1−s)
+
3s−2
,
with u solution to (1.1)-(1.2), d = 2. This bound implies the desired conclusion.
References
[1] J. Bourgain, Fourier transform restriction phenomena for certain lattice subsets and applications to
nonlinear evolution equations, GAFA 3 (1993), 107-156 and 209-262.
[2] J. Bourgain, Global solutions of nonlinear Schrödinger equations, AMS, Colloquium Publications 46
(1998).
[3] J. Bourgain, A remark on normal forms and the “I-method for periodic NLS, J. Anal. Math. 94 (2004),
125–157.
[4] J. Bourgain, Refinements of Strichartz’s inequality and applications to 2D NLS with critical nonlinearity, Intern. Mat. Res. Notices 5, (1998), 253-283.
[5] J. Bourgain, Remarks on stability and diffusion in high-dimensional Hamiltonian systems and partial
differential equations, Ergodic Theory Dynam. Systems 24 (2004), no. 5, 1331–1357.
[6] J. Bourgain, On Strichartz’ inequalities and the nonlinear Schrödinger equation on irrational tori,
Preprint.
[7] E. Bombieri and J. Pila, The number of integral points on arcs and ovals, Duke Math. J. 59 (1989),
no. 2, 337–357.
[8] T. Cazenave and F. Weissler, The Cauchy problem for the critical nonlinear Schröodinger equation in
H s , Nonlinear Analysis, Theory Methods and Applications, 14:10 (1990), 807-836.
[9] J. Colliander, J.-M. Delort, C. Kenig, G. Staffilani, Bilinear estimates and applications to 2D NLS,
Trans. Amer. Math. Soc. 353 (2001), no. 8, 3307–3325 (electronic).
[10] J. Colliander, M. Keel, G. Staffilani, H. Takaoka, T. Tao, Global Well-posedness for Schrödinger equations with derivative, SIAM J. Math. Anal. 33, no. 3, (2001), 649-669.
[11] J. Colliander, M. Keel, G. Staffilani, H. Takaoka, T. Tao, Sharp multi-linear periodic KdV estimates
and applications, J. Funct. Anal. 211 (2004), no. 1, 173-218.
[12] J. Colliander, M. Keel, G. Staffilani, H. Takaoka, T. Tao, A refined Global Well-posedness result for
Schrödinger equations with derivative, SIAM J. Math. Anal. 34 (2002) , 64-86.
[13] J. Colliander, M. Keel, G. Staffilani, H. Takaoka, T. Tao, Sharp global well-posedness for KdV and
modified KdV on R and T, J. Amer. Math. Soc. 16 (2003), 705-749.
[14] J. Colliander, M. Keel, G. Staffilani, H. Takaoka, T. Tao, Almost conservation laws and global rough
solutions to a nonlinear Schrödinger equation, Math. Research Letters 9 (2002), 659-682.
[15] J. Ginibre, Le problme de Cauchy pour des EDP semi-linaires priodiques en variables d’espace (d’aprs
Bourgain), Sminaire Bourbaki, Vol. 1994/95, Astrisque No. 237 (1996), Exp. No. 796, 4, 163–187.
[16] J. Ginibre and G. Velo, The global Cauchy problem for the nonlinear Schröodinger equation, H. Poincaré
Analyse Non Linéaire 2 (1985), 309-327.
[17] J. Ginibre, Y. Tsutsumi and G. Velo, On the Cauchy problem for the Zakharov system, J. Func. Anal.
151 (1997), 384-436
[18] A. Grünrock, On the Cauchy and periodic boundary value problem for a certain class of derivative
nonlinear Schrödinger equations, preprint.
[19] A. Iosevich, E. Sawyer and A. Seeger, Mean square discrepancy bounds for the number of lattice points
in large convex bodies J. Anal. Math. 87 (2002), 209–230.
[20] T. Kato, On nonlinear Schrödinger equations, Ann. Inst. H. Poincaré Physique Théorique 46 (1987),
113-129.
28
DANIELA DE SILVA, NATAŠA PAVLOVIĆ, GIGLIOLA STAFFILANI, AND NIKOLAOS TZIRAKIS
[21] C.E. Kenig, G. Ponce, and L. Vega, Quadratic forms for the 1-D semilinear Schrödinger equation,
Trans. Amer. Math. Soc. 348 (1996), No 8, 3323-3353.
[22] C.E. Kenig, G. Ponce and L. Vega, Well-posedness and scattering results for the generalized Korteveg-de
Vries equation via the contraction principle, Comm. Pure Appl. Math 46 (1993), 527–620.
[23] T. Ozawa and Y. Tsutsumi, Space-time estimates for null gauge forms and nonlinear Schrödinger
equations, Differential Integral Equations 11 (1998), 201-222.
[24] G Pick, Geometrisches zur Zahlenlehre, Sitzungber, Lotos (Prague) 19 (1899), 311-319.
[25] N. Tzirakis, The Cauchy problem for the semilinear quintic Schrödinger equation in one dimension,
Differential and Integral Equations 18 (2005), 947-960.
Department of Mathematics, Johns Hopkins University, Baltimore, MD 21218
E-mail address: desilva@math.jhu.edu
Department of Mathematics, Princeton University, Princeton, NJ 08544-1000
E-mail address: natasa@math.princeton.edu
Department of Mathematics, Massachusetts Institute of Technology, Cambridge, MA
02139-4307
E-mail address: gigliola@math.mit.edu
Department of Mathematics, University of Toronto, Toronto, Ontario, Canada M5S 2E4
E-mail address: tzirakis@math.toronto.edu